1. Field of the Invention
The present invention relates generally to the field of parametric loudspeakers. More particularly, this invention relates to the operation of parametric loudspeakers in a saturated air medium, or above and below saturation levels in the air medium while maintaining significantly reduced distortion.
2. Related Art
Audio reproduction has long been considered a well-developed technology. Over the decades, sound reproduction devices have moved from a mechanical needle on a tube or vinyl disk, to analog and digital reproduction over laser and many other forms of electronic media. Advanced computers and software now allow complex programming of signal processing and manipulation of synthesized sounds to create new dimensions of listening experience, including applications within movie and home theater systems. Computer generated audio is reaching new heights, creating sounds that are no longer limited to reality, but extend into the creative realms of imagination.
Nevertheless, the actual reproduction of sound at the interface of electro-mechanical speakers with the air has remained substantially the same in principle for almost one hundred years. Such speaker technology is clearly dominated by dynamic speakers, which constitute more than 90 percent of commercial speakers in use today. Indeed, the general class of audio reproduction devices referred to as dynamic speakers began with the simple combination of a magnet, voice coil and cone, driven by an electronic signal. The magnet and voice coil convert the variable voltage of the signal to mechanical displacement, representing a first stage within the dynamic speaker as a conventional multistage transducer. The attached cone provides a second stage of impedance matching between the electrical transducer and air envelope surrounding the transducer, enabling transmission of small vibrations of the voice coil to emerge as expansive compression waves that can fill an auditorium. Such multistage systems comprise the current fundamental approach to reproduction of sound, particularly at high energy levels.
A lesser category of speakers, referred to generally as film or diaphragmatic transducers, relies on movement of an emitter surface area of film that is typically generated by electrostatic or planar magnetic driver members. Although electrostatic speakers have been an integral part of the audio community for many decades, their popularity has been quite limited. Typically, such film emitters are known to be low-power output devices having limited applications. With a few exceptions, commercial film transducers have found primary acceptance as tweeters and other high frequency devices in which the width of the film emitter is equal to or less than the propagated wavelength of sound. Attempts to apply larger film devices have resulted in poor matching of resonant frequencies of the emitter with sound output, as well as a myriad of mechanical control problems such as maintenance of uniform spacing from the stator or driver, uniform application of electromotive fields, phase matching, frequency equalization, etc.
As with many well-developed technologies, advances in the state of the art of sound reproduction have generally been limited to minor enhancements and improvements within the basic fields of dynamic and electrostatic systems. Indeed, substantially all of these improvements operate within the same fundamental principles that have formed the basics of well-known audio reproduction. These include the concept that (i) sound is generated at a speaker face, (ii) based on reciprocating movement of a transducer (iii) at frequencies that directly stimulate the air into the desired audio vibrations. From this basic concept stems the myriad of speaker solutions addressing innumerable problems relating to the challenge of optimizing the transfer of energy from a dense speaker mass to the almost massless air medium that must propagate the sound.
A second fundamental principle common to prior art dynamic and electrostatic transducers is the fact that sound reproduction is based on a linear mode of operation. In other words, the physics of conventional sound generation rely on mathematics that conform to linear relationships between absorbed energy and the resulting wave propagation in the air medium. Such characteristics enable predictable processing of audio signals, with an expectation that a given energy input applied to a circuit or signal will yield a corresponding, proportional output when propagated as a sound wave from the transducer.
In such conventional systems, maintaining the air medium in a linear mode is extremely important. If the air is driven excessively into a nonlinear state, severe distortion occurs and the audio system is essentially unacceptable. This nonlinearity occurs when the air molecules adjacent the dynamic speaker cone or emitter diaphragm surface are driven to excessive energy levels that exceed the ability of the air molecules to respond in a corresponding manner to speaker movement. In simple terms, when the air molecules are unable to match the movement of the speaker so that the speaker is loading the air with more energy than the air can dissipate in a linear mode, then a nonlinear response occurs, leading to severe distortion and speaker inoperability. Conventional sound systems are therefore built to avoid this limitation, ensuring that the speaker transducer operates strictly within a linear range.
Parametric sound systems, however, represent an anomaly in audio sound generation. Instead of operating within the conventional linear mode, parametric sound can only be generated when the air medium is driven into a nonlinear state. Within this unique realm of operation, audio sound is not propagated from the speaker or transducer element. Instead, the transducer is used to propagate carrier waves of high-energy, ultrasonic bandwidth beyond human hearing. The ultrasonic wave therefore functions as the carrier wave, which can be modulated with audio input that develops sideband characteristics capable of decoupling in air when driven to the nonlinear condition. In this manner, it is the air molecules and not the speaker transducer that will generate the audio component of a parametric system. Specifically, it is the sideband component of the ultrasonic carrier wave that energizes the air molecule with audio signal, enabling eventual wave propagation at audio frequencies.
Another fundamental distinction of a parametric speaker system from that of conventional audio is that high-energy transducers as characterized in prior art audio systems do not appear to provide the necessary energy required for effective parametric speaker operation. For example, the dominant dynamic speaker category of conventional audio systems is well known for its high-energy output. Clearly, the capability of a cone/magnet transducer to transfer high-energy levels to surrounding air is evident from the fact that virtually all high-power audio speaker systems currently in use rely on dynamic speaker devices. In contrast, low output devices such as electrostatic and other diaphragm transducers are virtually unacceptable for high-power requirements. As an obvious example, consider the outdoor audio systems that service large concerts at stadiums and other outdoor venues. Normally, massive dynamic speakers are necessary to develop direct audio to such audiences. To suggest that a low-power film diaphragm might be applied in this setting would be considered foolish and impractical.
In summary, whereas conventional audio systems rely on well accepted acoustic principles of (i) generating audio waves at the face of the speaker transducer, (ii) based on a high-energy output device such as a dynamic speaker, (iii) while operating in a linear mode, the present inventors have discovered that just the opposite design criteria are preferred for parametric applications. Specifically, effective parametric sound is effectively generated using (i) a comparatively low-energy emitter, (ii) in a nonlinear mode, (iii) to propagate an ultrasonic carrier wave with a modulated sideband component that is decoupled in air (iv) at extended distances from the face of the transducer. In view of these distinctions, it is not surprising that much of the conventional wisdom developed over decades of research in conventional audio technology is simply inapplicable to problems associated with the generation parametric sound.
Despite developments in parametric sound, two main problems remain. First, is that parametric loudspeakers have historically only been capable of producing limited acoustic output. While it is clear that greater signal levels are needed, designers have historically limited the levels at which parametric speakers are driven in order to avoid driving the surrounding air medium into saturation. Saturation occurs where the air molecules are driven to such a high level of intensity, that they no longer accurately respond to the vibrations of the emitter. In prior parametric speakers, air saturation was avoided because high levels of distortion would typically result. Instead, parametric loudspeakers have required larger diameter, higher cost emitters to avoid saturating the air medium. While higher acoustic outputs and lower cost, smaller emitters are desirable in a parametric loudspeaker, these features have thus far been largely unattainable.
The second problem that still plagues parametric sound is that of reducing distortion levels at higher output levels. Based on the prior art Berktay solution, a reproduced audio frequency input signal should be distortion free where the signal has been square rooted before passing through double sideband AM modulation.
When the square-root processing is applied, testing by the inventors has shown that distortion is reduced only when the ultrasound level is small, and can increase dramatically with the ultrasound intensity. These data show that the prior art Berktay, square root preprocessing solution cannot effectively reduce distortion with high levels of ultrasound pressure. Furthermore, the square-root preprocessing method is not valid for a wide range of ultrasonic amplitudes, and particularly not valid for the higher intensity outputs required for improved parametric sound pressure levels. Finally, to perfectly reproduce the Berktay solution, an infinite number of terms are required, which is impractical to implement. It has been found that with square-root preprocessing, THD (Total Harmonic Distortion) can range between a few percent to as high as 50 percent or more as levels increase.
Accordingly, it would be an improvement over the current state of the art to provide a system of minimized size requirements that can provide higher acoustic output levels, while maintaining low distortion levels at all output levels.